cs.CL（2025-12-23）

📊 共 20 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (11 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗4) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (11 篇)

#	题目	一句话要点	标签	🔗	⭐
1	Can LLMs Solve My Grandma's Riddle? Evaluating Multilingual Large Language Models on Reasoning Traditional Bangla Tricky Riddles	BanglaRiddleEval：评估多语言大模型在孟加拉语传统谜语推理上的能力	large language model chain-of-thought	✅
2	Retrieval-augmented Prompt Learning for Pre-trained Foundation Models	提出RetroPrompt，通过检索增强提示学习提升预训练模型泛化能力	foundation model multimodal
3	M$^3$KG-RAG: Multi-hop Multimodal Knowledge Graph-enhanced Retrieval-Augmented Generation	提出M$^3$KG-RAG，通过多跳多模态知识图增强检索增强生成，提升MLLM在视听领域的推理和 grounding 能力。	large language model multimodal
4	Investigating Model Editing for Unlearning in Large Language Models	探索模型编辑算法用于大语言模型中的非学习，提升遗忘质量	large language model
5	Large Language Models Approach Expert Pedagogical Quality in Math Tutoring but Differ in Instructional and Linguistic Profiles	大型语言模型在数学辅导中接近专家级教学质量，但在教学和语言风格上存在差异	large language model
6	Making Large Language Models Efficient Dense Retrievers	提出EffiR框架，通过MLP压缩提升LLM密集检索器的效率，同时保持性能。	large language model
7	Cube Bench: A Benchmark for Spatial Visual Reasoning in MLLMs	提出Cube Bench：用于评估多模态大语言模型空间视觉推理能力的魔方基准测试。	large language model multimodal
8	EssayCBM: Rubric-Aligned Concept Bottleneck Models for Transparent Essay Grading	EssayCBM：一种基于规则对齐概念瓶颈模型的透明作文评分方法	large language model
9	Coherence in the brain unfolds across separable temporal regimes	利用LLM提取的漂移和位移信号，揭示大脑在自然听觉中连贯性处理的时域分离机制	large language model
10	AI Security Beyond Core Domains: Resume Screening as a Case Study of Adversarial Vulnerabilities in Specialized LLM Applications	揭示LLM在简历筛选中对抗性漏洞，提出FIDS防御机制	large language model
11	Schoenfeld's Anatomy of Mathematical Reasoning by Language Models	提出ThinkARM框架，解析语言模型数学推理过程中的认知结构与步骤	large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#	题目	一句话要点	标签	🔗	⭐
12	Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning	提出Nemotron 3 Nano，一种高效的混合专家Mamba-Transformer模型，用于Agent推理。	Mamba
13	FaithLens: Detecting and Explaining Faithfulness Hallucination	提出FaithLens，用于检测并解释大语言模型中的忠实性幻觉问题。	reinforcement learning large language model
14	Fun-Audio-Chat Technical Report	Fun-Audio-Chat：通过双分辨率语音表示和核心鸡尾酒训练，实现高效且强大的大型音频语言模型	DPO instruction following	✅
15	Multi-hop Reasoning via Early Knowledge Alignment	提出早期知识对齐(EKA)模块，提升迭代RAG多跳推理性能与效率	reinforcement learning large language model	✅
16	Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents	Memory-T1：利用强化学习进行多轮对话Agent中的时序推理	reinforcement learning	✅
17	Distilling to Hybrid Attention Models via KL-Guided Layer Selection	提出基于KL散度的层选择方法，用于将Softmax注意力Transformer蒸馏为混合注意力模型。	linear attention distillation
18	SpidR: Learning Fast and Stable Linguistic Units for Spoken Language Models Without Supervision	SpidR：一种无需监督即可学习快速稳定语音单元的语音语言模型	representation learning distillation	✅

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
19	Semantic Deception: When Reasoning Models Can't Compute an Addition	提出语义欺骗框架，揭示LLM在符号推理中易受语义误导的缺陷	manipulation large language model chain-of-thought
20	AprielGuard	提出AprielGuard，统一安全风险与对抗威胁，提升LLM安全防护能力	manipulation large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页